We should rename the protos here and here to flagd.evaluation.v1 and flagd.sync.v1, respectively.

Why?

gRPC services should be named/namespaced appropriately, particularly for the following reasons:

• collisions like this can occur when loading protos with the same name; not having a flagd namespace increases this likelihood significantly.
• when HTTP traffic is inspected, lack of flagd will make it difficult to identify, increasing likelihood of admins blocking traffic, etc
• in browser client usage, reverse proxy rules for /schema requests aren't obviously related to flagd; improving the name will make such rules' purpose obvious and prevent clashes
• semantic names are generally important, and schema.v1 doesn't say anything about what this RPC/service actually DOES.

How?

In order to reduce breakage, we should support both the old and new RPCs. I'm not sure how best to do this. There may be a way to "proxy" traffic from one to the other. Alternatively, we might have to build and run both protos (so we would pull 2 versions of the protobuff schema).

We would then slowly update providers to use the new protos, and deprecate the old proto in flagd 1.0.